On October 2, 2026, Capcom announced a multi-stage overhaul of its proprietary RE Engine, dubbed Project REX, during the Capcom Open Conference RE: 2026 tech event in Japan. The initiative aims to transform the foundation of Capcom's flagship franchises into what the company calls an "AI-generation game engine," preparing internal studios for a future where developers build games directly alongside artificial intelligence.
Capcom programmer Satoshi Ishida detailed the technological roadmap in a keynote titled "The Outlook and Future of the REX Project, Further Evolving the RE Engine for the Next Generation". Rather than abandoning the battle-tested framework and building a new engine from the ground up, Capcom is modularly upgrading its existing architecture. The primary mission is addressing ballooning production times, managing massive asset pipelines, and directly embedding machine learning into internal workflows.
What Capcom Announced at RE: 2026
During the keynote presentation, Ishida identified three mounting challenges that made a fundamental engine upgrade unavoidable:
- Escalating game scale: As modern triple-A releases grow exponentially in asset volume, simple editing, compiling, and data-management tasks have become bottlenecked and time-consuming.
- A massive, international workforce: RE Engine is now used daily by more than 2,000 developers across diverse technical disciplines and nationalities, creating an urgent need for standardized, accessible toolchains.
- Essential AI integration: Artificial intelligence has advanced to the point where integrating it directly into the development pipeline—rather than bolting it on as an external patch—is a core requirement to keep modern game production viable.
To solve these structural friction points, Capcom is building Project REX around the philosophy of creating games "together with AI". By unifying the engine's underpinnings around the world's most ubiquitous programming standards, Capcom intends to make the codebase equally readable to human software engineers and large-scale AI models. Under this structure, Capcom expects machine-learning models to assist engineers with writing routine code, compiling data, and conducting automated playtests to identify bugs long before manual QA teams step in.
Watch: RE:2023 RE ENGINE's Past and Future
Video: CAPCOM R&D on YouTube
Inside Project REX: The Six Core Pillars
Photo: The Verge (source)
Capcom is deploying Project REX across six distinct modules, each designed to overhaul specific pain points in data handling, interface performance, execution speed, and automated testing.
| REX Component | Primary Role | AI and Workflow Benefit |
|---|---|---|
| Re:Dox | Data standardizer | Translates heterogeneous data formats into unified rule sets for rapid processing |
| RE:UI | Tool interface layer | Separates visual interface from internal state, speeding up internal editors |
| RE:Log | Diagnostic logging | Replaces heavy text logs with lightweight, ID-based tracking for instant debugging |
| RE:Flows | Visual scripting | Converts node-based visual scripts into optimized native code readable by AI |
| RE:Runtime | Scene execution core | Uses batch processing to handle scenes with massive crowds and complex entities |
| RE:C++ | Engine execution layer | Unifies the native coding layer for maximum compilation speed and AI syntax parsing |
These components solve very specific internal bottlenecks. For instance, Re:Dox acts as a universal translator, standardizing game parameters so AI algorithms can parse project data without choking on mismatched file formatting. Meanwhile, RE:UI drastically reduces latency in editor tools—a change developers likened to adding hardware acceleration to heavy creation suites—ensuring tools remain responsive even when handling tens of gigabytes of world geometry.
On the simulation side, RE:Runtime re-architects the core game loop using batch-processing techniques. This allows future games to render and simulate significantly denser crowds and dynamic NPCs simultaneously without draining console and PC CPU threads. To support these advancements, Capcom has also open-sourced select underlying elements of the technology to foster broader algorithmic training.
The History: From Resident Evil 7 to Modern Flagships
Capcom first kicked off development on RE Engine in 2014, designing it to replace its aging MT Framework. The engine made its public debut with the release of Resident Evil 7: Biohazard in 2017, earning immediate industry praise for its photorealistic PBR (physically based rendering) lighting, robust performance, and rapid load times.
Over the following decade, the engine proved remarkably versatile. What began as an intimate first-person survival horror engine expanded to power nearly 30 shipped titles spanning drastically different genres:
- Action & Fighting: Devil May Cry 5, Street Fighter 6
- Expansive RPGs: Monster Hunter Rise, Dragon's Dogma 2
- Remakes & Sequels: Resident Evil 2, 3, and 4 Remakes, Resident Evil Village
- New Frontiers: Exoprimal, Kunitsu-Gami: Path of the Goddess, Pragmata
However, expanding an engine initially tailored for tight, claustrophobic corridors into vast open-world landscapes like Dragon's Dogma 2 revealed technical limitations. CPU-heavy simulations and dense character counts taxed development pipelines. Project REX is Capcom’s long-term answer to those growing pains, systematically modernizing the engine without throwing away a decade of specialized internal tool familiarity.
Proprietary Engineering vs. Third-Party Engines
Photo: ign.com (source)
Capcom’s heavy investment in Project REX reinforces a clear industry divide. While major publishers such as CD Projekt Red, EA, and Square Enix have increasingly abandoned proprietary tech in favor of Epic Games' Unreal Engine 5, Capcom remains committed to in-house engine ownership.
Maintaining a proprietary engine allows Capcom to avoid external licensing fees—which typically take up to 5% of gross game revenue with commercial engines like Unreal—while tailoring rendering pipelines strictly to its internal design paradigms. However, modern engines like Unreal Engine 5 and Unity 6 are also aggressively rolling out machine-learning suites and neural rendering techniques.
Capcom's approach differs fundamentally from turnkey engines. By building REX around common C++ standards, Capcom is creating a closed, proprietary pipeline that can plug directly into modern machine-learning models without exposing confidential intellectual property to external cloud networks.
Industry and Community Reactions
The announcement triggered immediate discussion across development circles and gaming forums. Gamers and digital artists have grown wary of publishers chasing artificial intelligence, frequently pointing to poorly received experiments with generative art, algorithmic voice cloning, and executive cost-cutting across Western studios.
However, technical analysts quickly pointed out the critical distinction between Capcom’s REX initiative and generic "generative AI" hype. Capcom's presentation focused strictly on procedural workflow acceleration, code comprehension, and automated QA bug hunting rather than generating creative assets out of thin air.
Capcom has previously stated that it does not intend to place generative AI art or audio into its consumer releases. Community sentiment largely echoed cautious optimism: players voiced hope that using AI for mundane bug hunting and data translation will allow human creative teams to focus on bespoke game design and polished launch performance.
What Project REX Means for Gamers
Photo: eurogamer.net (source)
For consumer players, Capcom's REX roadmap offers several tangible long-term benefits:
- Shorter gaps between major releases: Streamlining compilation pipelines and asset conversion prevents projects from getting stuck in multi-year production bottlenecks.
- More stable day-one launches: Using machine-learning agents to run thousands of automated, continuous playtests will catch game-breaking physics glitches and performance regressions earlier in production.
- Better PC and console optimization: REX’s overhauled runtime and batch-processing architecture are directly engineered to prevent CPU bottlenecks when dozens or hundreds of NPCs appear on screen simultaneously.
- Consistent creative direction: Because the engine focuses on engineering productivity rather than replacing creative staff, Capcom's signature visual fidelity and art direction remain human-driven.
What’s Next and Unanswered Questions
Because REX represents an ongoing, phased transformation rather than a hard engine break, Capcom will roll out these six modules across its internal production pipelines over several production cycles. The company has not yet confirmed which upcoming title will serve as the first official showcase for the completed REX feature set.
Key questions remain regarding how deeply AI-assisted scripting will change everyday programming roles inside Capcom's studios. Observers are also watching closely to see whether Capcom's open-source components will lead to wider industry collaboration with external academic and machine-learning research labs. As production scales into the late 2020s, Capcom's pragmatic approach to AI may well set the benchmark for how triple-A studios harness automated tools without sacrificing creative integrity.
FAQ
Is Capcom replacing game developers and artists with generative AI? No. Capcom has stated that it does not intend to use AI to generate game art or assets wholesale, focusing instead on internal developer tools, data standardizing, and automated code testing.
What does the name REX stand for? REX is the internal codename for the next-generation evolution of Capcom's proprietary RE Engine, adding modular upgrades rather than creating a completely new engine from scratch.
When was Capcom's original RE Engine created? Work on RE Engine began internally in 2014, and it debuted commercially in January 2017 with Resident Evil 7: Biohazard.
How will AI be used during game development under Project REX? Capcom is structuring the engine's codebase around widely understood programming languages so AI can assist engineers with generating routine code, converting data formats, and running automated playtests to detect bugs.
Which games currently run on RE Engine? Nearly 30 titles utilize RE Engine, including recent entries in the Resident Evil franchise, Street Fighter 6, Devil May Cry 5, Monster Hunter Rise, and Dragon's Dogma 2.
